Insecticidal Protein. Bacillus thuringiensis, a species of bacteria closely related to B. cereus, forms a proteinaceous crystalline inclusion during sporulation. This crystal is parasporal, forming within the cell at the end opposite from the developing spore. The crystal protein, often referred to as the δ-endotoxin, has two forms: a nontoxic protoxin of approximate molecular weight (MW) of 130 kilodaltons (kD), and a toxin having an approximate MW of 67 kD. The crystal contains the protoxin protein which is activated in the gut of larvae of a number of insect species. Klowden, M. J. et al. (1983) Appl. Environ. Microbiol. 46:312-315, have shown solubilized protoxin from B. thuringiensis var. israelensis is toxic to Aedes aegypti adults. During activation, the protoxin is cleaved into two polypeptides, one or both of which are toxic. In vivo, the crystal is activated by being solubilized and converted to toxic form by the alkalinity and proteases of the gut. In vivo the protoxin may be solubilized by extremely high pH (e.g., pH 12), by reducing agents under moderately basic conditions (e.g., pH 10), or by strong denaturants (guanidium, urea) under neutral conditions (pH 7), and once solubilized, may be activated by the action of the protease trypsin. The crystal protein is reported to be antigenically related to proteins within both the spore coat and the vegetative cell wall. Carbohydrate is not involved in the toxic properties of the protein.
B. thuringiensis and its crystalline endotoxin are useful because the crystal protein is an insecticidal protein known to be poisonous to the larvae of over a hundred species of insects, most commonly those from the orders Lepidoptera and Diptera. Insects susceptible to the action of the B. thuringiensis crystal protein include, but need not be limited to, those listed in Table 1. Many of these insect species are economically important pests. Plants which can be protected by application of the crystal protein include, but need not be limited to, those listed in Table 2. Different varieties of B. thuringiensis, which include, but need not be limited to, those listed in Table 3, have different host ranges (Faust, R. M. et al. [1982] in Genetic Engineering in the Plant Sciences, Panapolous, N.J. (ed.), pp. 225-254); this probably reflects the toxicity of a given crystal protein in a particular host. The crystal protein is highly specific to insects; in over two decades of commercial application of sporulated B. thuringiensis cells to crops and ornamentals, there has been no known case of effects to plants or non-insect animals. The efficacy and safety of the endotoxin have been reviewed by Faust, R. M. et al., supra. Other useful reviews include those by Fast, P. G. (1981) in Microbial Control of Pests and Plant Diseases, 1970-1980, Burges, H. D. (ed.), pp. 223-248; and Huber, H. E., P. Luthy (1981) in Pathogenesis of Invertebrate Microbial Diseases, Davidson, E. W. (ed.), pp. 209-234.
The crystal protein gene usually can be found on one of several large plasmids that have been found in Bacillus thuringiensis, though in some strains it may be located on the chromosome (Kronstad, J. W. et al. [1983] J. Bacteriol. 154:419-428). Several of the genes have been cloned into plasmids that can grow in E. coli Whiteley's group (Whiteley, H. R. et al. [1982] in Molecular Cloning and Gene Regulation in Bacilli, Ganesan, A. T. et al. (eds.), pp. 131-144; Schnepf, H. E., H. R. Whiteley [1981] Proc. Natl. Acad. Sci USA 78:2893-2897; and European patent application 63,949) reported the cloning of the toxin from B. thuringiensis var. kurstaki strains HD-1-Dipel and HD-73, using the enzymes Sau3AI (under partial digest conditions) and BglII, respectively, to insert large gene-bearing fragments having approximate sizes of 12 kilobasepairs (kbp) and 16 kbp into the BamHI site of the E. coli plasmid vector pBR322. The HD-1 crystal protein was observed to be located on a 6.6 kbp HindIII fragment. Crystal protein from the HD-1-Dipel gene which was toxic to larvae, immunologically identifiable, and the same size as the authentic protoxin, was observed to be produced by transformed E. coli cells containing pBR322 clones or subclones. This indicated that the Bacillus gene was transcribed, probably from its own promoter, and translated in E. coli. Additionally, this suggests that the toxic activity of the protein product is independent of the location of its synthesis. That the gene was expressed when the fragment containing it was inserted into the vector in either orientation suggests that transcription was controlled by its own promoter. The transcriptional and translational start sites, as well as the deduced sequence for the amino-terminal 333 amino acids of the HD-1-Dipel protoxin, have been determined by nucleic acid sequence (Wong, H. C. et al. [1983] J. Biol. Chem. 258:1960-1967). The insecticidal gene was found to have the expected bacterial ribosome binding and translational start (ATG) sites along with commonly found sequences at −10 and −35 (relative to the 5′-end of the mRNA) that are involved in initiation of transcription in bacteria such as B. subtilis. Klier, A. et al. (1982) EMBO J. 1:791-799, have reported the cloning of the crystal protein gene from B. thuringiensis strain berliner 1715 in pBR322. Using the enzyme BamHI, a large 14 kbp fragment carrying the crystal protein gene was moved into the vector pHV33, which can replicate in both E. coli and Bacillus. In both E. coli and sporulating B. subtilis, the pHV33-based clone directed the synthesis of full-size (130 kD) protoxin which formed cytoplasmic inclusion bodies and reacted with antibodies prepared against authentic protoxin. Extracts of E. coli cells harboring the pBR322 or pHV33-based plasmids were toxic to larvae. In further work, Klier, A et al. (1983) Nucl. Acids Res. 11:3973-3987, have transcribed the berliner crystal protein gene in vitro and have reported on the sequence of the promoter region, together with the first 11 codons of the crystal protein. The bacterial ribosome binding and translational start sites were identified. Though the expected “−10” sequence was identified, no homology to other promoters has yet been seen near −35. Held et al. (1982) Proc. Natl. Acad. Sci. USA 77:6065-6069, reported the cloning of a crystal protein gene from the variety kurstaki in the phage g-based cloning vector Charon4A E. coli cells infected with one of the Charon clones produced antigen that was the same size as the protoxin (130 kD) and was toxic to larvae. A 4.6 kbp EcoRI fragment of this Charon clone was moved into pHV33 and an E. coli plasmid vector, pBR328. Again, 130 kD antigenically identifiable crystal protein was produced by both E. coli and B. subtilis strains harboring the appropriate plasmids. A B. thuringiensis chromosomal sequence which cross-hybridized with the cloned crystal protein gene was identified in B. thuringiensis strains which do not produce crystal protein during sporulation.
In addition to the crystal protein, B. thuringiensis produces at least three other toxins. Two of them, the α-exotoxin and γ-exotoxin, are phospholipase elements that degrade lipids. B. cereus is also known to produce phospholipases (or lecithinases) which are toxic to insect larvae. Other bacterial enzymes which are involved in insect pathogenesis include, but need not be limited to, hyaluronidases, phosphatases, and proteases. Protease produced by Pseudomonas aeruginosa has been shown to have a specific affinity to proteins of Galleria mellonella larvae (see Lysenko, O., M. Kucera [1971] in Microbial Control of Insects and Mites, Burges, H. D., N. W. Hussey (eds.), pp. 205-277).
Chang, S. (1983) Trends Biotechnol. 1:100-101, reported that the DNA sequence of a complete HD-1 gene had been publicly presented (ref. 5 therein), and that the HD-1 toxin moiety resides in the amino-terminal 67 kD peptide.
Shuttle Vectors. Shuttle vectors, developed by Ruvidn, G. B., F. M. Ausubel (1981) Nature 298:85-88, provide a way to insert foreign genetic materials into position of choice in a large plasmid, virus, or genome. There are two main problems encountered when dealing with large plasmids or genomes. First, the large plasmids may have many sites for each restriction enzyme. Unique site-specific cleavage reactions are not reproducible, and multi-site cleavage reactions followed by ligation lead to great difficulties due to the scrambling of the many fragments whose order and orientation one does not want changed. Second, the transformation efficiency with large DNA plasmids is very low. Shuttle vectors allow one to overcome these difficulties by facilitating the insertion, often in vitro, of the foreign genetic material into a smaller plasmid, then transferring, usually by in vivo techniques, to the larger plasmid.
A shuttle vector consists of a DNA molecule, usually a plasmid, capable of being introduced into the ultimate recipient bacteria. It also includes a copy of the fragment of the recipient genome into which the foreign genetic material is to be inserted and a DNA segment coding for a selectable trait, which is also inserted into the recipient genome fragment. The selectable trait (“marker”) is conveniently inserted by transposon mutagenesis or by restriction enzymes and ligases.
The shuttle vector can be introduced into the ultimate recipient cell, typically a bacterium of the family Rhizobiaceae (which contains the genus Agrobacterium), by a tri-parental mating (Ruvkin and Ausubel, supra), direct transfer of a self-mobilizable vector in a bi-parental mating, direct uptake of exogenous DNA by Agrobacterium cells (“transformation,” using the conditions of Holsters, M. et al. [1978] Mol. Gen. Genet. 163:181-187), by spheroplast fusion of Agrobacterium with another bacterial cell, by uptake of liposome-encapsulated DNA, or by infection with a shuttle vector that is based on a virus that is capable of being packaged in vitro. A tri-parental mating involves the mating of a strain containing a mobilizable plasmid, which carries genes for plasmid mobilization and conjugative transfer, with the strain containing the shuttle vector. If the shuttle vector is capable of being mobilized by the plasmid genes, the shuttle vector is transferred to the recipient cell containing the large genome, e.g., the Ti or Ri plasmids of Agrobacterium strains.
After the shuttle vector is introduced into the recipient cell, possible events include a double crossover with one recombinational event on either side of the marker. This event will result in transfer of a DNA segment containing the marker to the recipient genome replacing a homologous segment lacking the insert. To select for cells that have lost the original shuttle vector, the shuttle vector must be incapable of replicating in the ultimate host cell or be incompatible with an independently selectable plasmid pre-existing in the recipient cell. One common means of arranging this is to provide in the third parent another plasmid which is incompatible with the shuttle vector and which carries a different drug resistance market. Therefore, when one selects for resistance to both drugs, the only surviving cells are those in which the marker on the shuttle vector has recombined with the recipient genome. If the shuttle vector carries an extra marker, one can then screen for and discard cells that contain plasmids resulting from a single crossover event between the shuttle vector and the recipient plasmid resulting in cointegrates in which the entire shuttle vector is integrated with the recipient plasmid. If the foreign genetic material is inserted into or adjacent to the marker that is selected for, it will also be integrated into the recipient plasmid as a result of the same double recombination. It might also be carried along when inserted into the homologous fragment at a spot not within or adjacent to the marker, but the greater the distance separating the foreign genetic material from the marker, the more likely will be a recombinational event occurring between the foreign genetic material and marker, preventing transfer of the foreign genetic material.
If the shuttle vector is used to introduce a phenotypically dominant trait (e.g., a novel expressible insecticide structural gene, but not an inactivated oncogenic T-DNA gene), one need not only rely on a double homologous recombination. The cells resulting from a single cross-over event resulting in cointegrate plasmids can transfer the desired trait into plant cells. One may even use a variant shuttle vector having a single uninterrupted sequence of T-DNA. However, as the resulting T-DNA will now contain a tandem duplication, one must be vigilant regarding a possible rare deletion of the shuttle vector by a single homologous recombination event occurring between the two homologous sequences in either the Agrobacterium or plant cells.
Shuttle vectors have proved useful in manipulation of Agrobacterium plasmids: see Garfinkel, D. J. et al. (1981) Cell 27:143-153; Matzke, A. J. M., M.-D. Chilton (1981) J. Mol. Appl. Genet. 1:39-49; and Leemans, J. et al. (1981) J. Mol. Appl. Genet. 1:149-164, who referred to shuttle vectors by the term “intermediate vectors.”
A recently disclosed variation of the shuttle vector system for inserting changes into large DNA molecules is the “suicide vector.” In this system, as described by Puhler, A. et al., U.S. Pat. No. 4,626,504, and Simon, R. et al. (November 1983) “A Broad Host Range Mobilization System for In Vivo Genetic Engineering: Transposon Mutagenesis in Gram-Negative Bacteria,” Biotechnology, pp. 784-791, the shuttle vector is incapable of being maintained within the recipient cell. This property eliminates the need to introduce an incompatible plasmid into the recipient cell in order to exclude the shuttle vector as is commonly done during a triparental mating. All vectors which do not integrate into some already present DNA effectively “commit suicide” by not being replicated. As can be done with traditional types of shuttle vectors, one may distinguish between double and single homologous recombination by screening for an antibiotic resistance gene which is not between the two regions of homology. Use of a pBR322-based suicide vector to transfer DNA sequences into a Ti plasmid has been reported by Van Haute, E. et al. (1983) EMBO J. 2:411-417, and Comai, L. et al. (1982) Plant Mol. Biol. 1:291-300.
An alternative to the use of shuttle vectors for introduction of novel DNA sequences into T-DNA by means of homologous recombination involves bacterial transposons. As described in the section Agrobacterium-Genes on the TIP Plasmids, transposons can “jump” into the T-DNA of a TIP plasmid (e.g., see Garfinkel, D. J. et al. [1981] Cell 27:143-153). Should the transposon be modified in vitro by the insertion of the novel sequence, that novel DNA can be transferred into the TIP plasmid's T-DNA by the transposon. The TIP can then transfer the novel DNA/transposon/T-DNA combination to a plant cell where it will be stably integrated.
Agrobacterium —Overview. Included within the Gram-negative bacterial family Rhizobiaceae in the genus Agrobacterium are the species A. tumefaciens and A. rhizogenes. These species are respectively the causal agents of crown gall disease and hairy root disease of plants. Crown gall is characterized by the growth of a gall of dedifferentiated tissue. Hairy root is a teratoma characterized by inappropriate induction of roots in infected tissue. In both diseases, the inappropriately growing plant tissue usually produces one or more amino acid derivatives, known as opines, not normally produced by the plant, which are catabolized by the infecting bacteria. Known opines have been classified into three main families whose type members are octopine, nopaline, and agropine. The cells of inappropriately growing tissues can be grown in culture and, under appropriate conditions, be regenerated into whole plants that retain certain transformed phenotypes.
Virulent strains of Agrobacterium harbor large plasmids known as Ti (tumor-inducing) plasmids in A. tumefaciens and Ri (root-inducing) plasmids in A. rhizogenes. Curing a strain of these plasmids results in a loss of pathogenicity. The Ti plasmid contains a region, referred to as T-DNA (transferred-DNA), which in tumors is found to be integrated into the genome of the host plant. The T-DNA encodes several transcripts. Mutational studies have shown that some of these are involved in induction of tumorous growth. Mutants in the genes for tml, tmr, tms, respectively, result in large tumors (in tobacco), a propensity to generate roots, and a tendency for shoot induction. The T-DNA also encodes the gene for at least one opine synthase, and the Ti plasmids are often classified by the opine which they caused to be synthesized. Each of the T-DNA genes is under control of a T-DNA promoter. The T-DNA promoters resemble eukaryotic promoters in structure, and they appear to function only in the transformed plant cell. The Ti plasmid also carries genes outside the T-DNA region. These genes are involved in functions which include opine catabolism, oncogenicity, agrocin sensitivity, replication, and autotransfer to bacterial cells. The Ri plasmid is organized in a fashion analogous to the Ti plasmid. The set of genes and DNA sequences responsible for transforming the plant cell are hereinafter collectively referred to as the transformation-inducing principle (TIP). The designation TIP therefore includes both Ti and Ri plasmids. The integrated segment of a TIP is termed herein “T-DNA” (transferred DNA), whether derived from a Ti plasmid or a Ri plasmid.
Chilton, M.-D. (June 1983) Sci. Amer. 248(6):50-59, has recently provided an introductory article on the use of Ti plasmids as vectors. Recent general reviews of Agrobacterium-caused disease include those by Merlo, D. J. (1982) Adv. Plant Pathol. 1:139-178; Ream, L. W., Gordon, M. P. (1982) Science 218:854-859; Bevan, M. W.; M.-D. Chilton (1982) Ann. Rev. Genet. 16:357-384; Kahl, G., J. Schnell (1982) Molecular Biology of Plant Tumors; and Barton, K. A., M.-D. Chilton (1983) Methods Enzymol. 101:527-539.
Agrobacterium —Infection of Plant Tissues. Plant cells can be transformed by Agrobacterium in a number of methods known in the art which include, but are not limited to, co-cultivation of plant cells in culture with Agrobacterium, direct infection of a plant, fusion of plant protoplasts with Agrobacterium spheroplasts, direct transformation by uptake of free DNA by plant cell protoplasts, transformation of protoplasts having partly regenerated cell walls with intact bacteria, transformation of protoplasts by liposomes containing T-DNA, use of a virus to carry in the T-DNA, microinjection, and the like. Any method will suffice as long as the gene is reliably expressed and is stably transmitted through mitosis and meiosis.
The infection of plant tissue by Agrobacterium is a simple technique well known to those skilled in the art (for an example, see Butcher, D. N. et al. [1980] in Tissue Culture Methods for Plant Pathologists, Ingram, D. S., J. P. Helgeson (eds.), pp. 203-208). Typically, a plant is wounded by any of a number of ways, which include cutting with a razor, puncturing with a needle, or rubbing with an abrasive. The wound is then inoculated with a solution containing tumor-inducing bacteria. An alternative to the infection of intact plants is the inoculation of pieces of tissues such as potato tuber discs (Anand, D. K., G. T Heberlein [1977] Amer. J. Both. 64:153:158) or segments of tobacco stems (Barton, K. A. et al. [1983] Cell 32:1033-1043. After induction, the tumors can be placed in tissue culture on media lacking phytohormones. Hormone-independent growth is typical of transformed plant tissue and is in great contrast to the usual conditions of growth of such tissue in culture (Braun, A. C. [1956] Cancer Res. 16:53-56).
Agrobacterium is also capable of infecting isolated cells and cells grown in culture (Marton, L. et al. [1979] Nature 277:129-131) and isolated tobacco mesophyll protoplasts. In the latter technique, after allowing time for partial regeneration of new cell walls, Agrobacterium cells were added to the culture for a time and then killed by the addition of antibiotics. Only those cells exposed to A. tumefaciens cells harboring the Ti plasmid were capable of forming calli when plated on media lacking hormone. Most calli were found to contain an enzymatic activity involved in opine anabolism. Other workers (Horsch, R. B., R. T. Fraley [18 Jan. 1983] 15th Miami Winter Symposium) have reported transformations by co-cultivation, leading to a high rate (greater than 10%) of calli displaying hormone-independent growth, with 95% of those calli making opines. Davey, M. R. et al. (1980) in Ingram and Helgeson, supra, pp. 209-219, describe the infection of older cells that had been regenerated from protoplasts.
Plant protoplasts can be transformed by the direct uptake of TIP plasmids. Davey, M. R. et al. (1980) Plant Sci Lett. 18:307-313, and Davey, M. R. et al. (1980) in Ingram and Helgeson, supra, were able to transform Petunia protoplasts with the Ti plasmid in the presence of poly-L-α-ornithine to a phenotype of opine synthesis and hormone-independent growth in culture. It was later shown (Draper, J. et al. [1982] Plant and Cell Physiol 23:451-458; Davey, M. R. et al. [1982] in Plant Tissue Culture, Fujiwara, A. (ed.), pp. 515-516) that polyethylene glycol stimulated Ti plasmid uptake and that some T-DNA sequences were integrated into the genome. Krens, F. A. et al. (1982) Nature 296:72-74, reported similar results using polyethylene glycol followed by a calcium shock, though their data suggests that the integrated T-DNA included flanking Ti plasmid sequences.
An alternative method to obtain DNA uptake involves the use of liposomes. The preparation of DNA containing liposomes is taught by Papahadjopoulos in U.S. Pat. Nos. 4,078,052 and 4,235,871. Preparations for the introduction of Ti-DNA via liposomes have been reported (Nagata, T. et al. [1982] in Fujiwara, supra, pp. 509-510; and Nagata, T. [1981] Mol. Gen. Genet. 184:161-165). An analogous system involves the fusion of plant and bacterial cells after removal of their cell walls. An example of this technique is the transformation of Vinca protoplast by Agrobacterium spheroplasts reported by Hasezawa, S. et al. (1981) Mol. Gen. Genet. 182:206-210. Plant protoplasts can take up cell wall delimited Agrobacterium cells (Hasezawa, S. et al. [1982] in Fujiwara, supra, pp. 517-518).
T-DNA can be transmitted to tissue regenerated from a fusion of two protoplasts, only one of which had been transformed (Wullens, G. J. et al. [1980] Theor. Appl. Genet. 56:203-208). As detailed in the section on Regeneration of Plants, T-DNA can pass through meiosis and be transmitted to progeny as a simple Mendelian trait.
Agrobacterium —Regeneration of Plants. Differentiated plant tissues with normal morphology have been obtained from crown gall tumors. Braun, A. C., H. N. Wood (1976) Proc. Natl. Acad. Sci. USA 73:496-500, grafted tobacco teratomas onto normal plants and were able to obtain normally appearing shoots which could flower. The shoots retained the ability to make opines and to grow independently of phytohormones when placed in culture. In the plants screened, these tumorous phenotypes were not observed to be transmitted to progeny, apparently being lost during meiosis (Turgeon, R. et al. [1976] Proc. Natl. Acad. Sci. USA 73:3562-3564). Plants which had spontaneously lost tumorous properties, or which were derived from teratoma seed, were initially shown to have lost all their T-DNA (Yang, F.-M. et al. [1980] In Vitro 16:87-92; Yang, F. et al. [1980] Mol. Gen. Genet. 177:707-714; Lemmers, M. et al. [1980] J. Mol. Biol. 144:353-376). However, later work with plants that had become revertants after hormone treatment (1 mg/1 kinetin) showed that plants that had gone through meiosis, though losing T-DNA genes responsible for the transformed phenotype, could retain sequences homologous to both ends of T-DNA (Yang, F., R. B. Simpson [1981] Proc. Natl. Acad. Sci. USA 78:4151-4155). Wullems, G. J. et al. [1981] Cell 24:719-724, further demonstrated that genes involved in opine anabolism were capable of passing through meiosis though the plants were male sterile and that seemingly unaltered T-DNA could be inherited in a Mendelian fashion (Wullems, G. J. [1982] in Fujiwara, supra). Otten, L. et al. (1981) Mol. Gen. Genet. 183:209-213, used Tn7 transposon-generated Ti plasmid mutants in the tms (shoot-inducing) locus to create tumors which proliferated shoots. When these shoots were regenerated into plants, they were found to form self-fertile flowers. The resultant seeds germinated into plants which contained T-DNA and made opines. In further experiments, DeGreve, H. et al. (1982) Nature 300:752-755, have found that octopine synthase can be inherited as a single dominant Mendelian gene. However, the T-DNA had sustained extensive deletions of functions other than ocs while undergoing regeneration from callus. Similar experiments with a tmr (root-inducing) mutant showed that full-length T-DNA could be transmitted through meiosis to progeny, that in those progeny nopaline genes could be expressed, though at variable levels, and that co-transformed yeast alcohol dehydrogenase I gene was not expressed (Barton, K. A. et al. [1983] Cell 32:1033-1043. It now appears that regenerated tissues which lack T-DNA sequences are probably descended from untransformed cells which “contaminate” the tumor (Ooms, G. et al. [1982] Cell 30:589-597). Recent work by Binns, A. N. (1983) Planta 158:272-279, indicates that tumorigenic genes, in this case tmr, can be “shut off” during regeneration and “turned back on” by placing regenerated tissue in culture.
Roots resulting from transformation from A. rhizogenes have proven relatively easy to regenerate directly into plantlets (Chilton, M.-D. et al. [1982] Nature 295:432-434).
Agrobacterium —Genes on the TIP Plasmids. A number of genes have been identified within the T-DNA of the TIP plasmids. About half a dozen octopine plasmid T-DNA transcripts have been mapped (Gelvin, S. B. et al. [1982] Proc. Natl. Acad. Sci. USA 79:76-80; Willmitzer, L. et al. [1982] EMBO J. 1:139-146) and some functions have been assigned (Leemans, J. et al. [1982] EMBO J. 1:147-152). Some of these transcripts, specifically those in the region encoding tmr and tms, can also be transcribed in prokaryotic cells (Schroder, G. et al. [1983] EMBO J. 2:403-409). The four genes of an octopine type plasmid that have been well defined by transposon mutagenesis include tms, tmr, and tml (Garfinkel, D. J. et al. [1981] Cell 27:143-153). Ti plasmids which carry mutations in these genes respectively incite tumorous calli of Nicotiana tabacum which generate shoots, proliferate roots, and are larger than normal. In other hosts, mutants of these genes can induce different phenotypes (see Bevan, M. W., M.-D. Chilton [1982] Ann. Rev. Genet. 16:357-384). The phenotypes of tms and tmr are correlated with differences in the phytohormone levels present in the tumor. The differences in cytokinin:auxin ratios are similar to those which in culture induce shoot or root formation in untransformed callus tissue (Aidyoshi, D. E. et al. [1983] Proc. Natl. Acad. Sci. USA 80:407-411). T-DNA containing a functional gene for either tms or tmr alone, but not functional tml alone, can promote significant tumor growth. Promotion of shoots and roots is respectively stimulated and inhibited by functional tml (Ream, L. W. et al. [1983] Proc. Natl. Acad. Sci. USA 80:1660-1664). Mutations of T-DNA genes do not seem to affect the insertion of T-DNA into the plant genome (Leemans et al. [1982] supra; Ream et al. [1983] supra). The ocs gene encodes octopine synthase, which has been sequenced by De Greve, H. et al. (1982) J. Mol. Appl. Genet. 1:499-511. It does not contain introns (intervening sequences commonly found in eukaryotic genes which are post-transcriptionally spliced out of the messenger precursor during maturation of the mRNA). It does have sequences that resemble a eukaryotic transcriptional signal (“TATA box”) and a polyadenylation site. All of the signals necessary for expression of the ocs gene are found within 295 bp of the ocs transcriptional start site (Koncz, C. et al. [1983] EMBO J. 2:1597-1603).
Nopaline Ti plasmids encode the nopaline synthase gene (nos), which has been sequenced by Depicker, A et al. (1982) J. Mol. Appl. Genet. 1:561-573. As was found with the ocs gene, nos is not interrupted by introns. It has two putative polyadenylation sites and a potential “TATA box.” In contrast to ocs, nos is preceded by a sequence which may be a transcriptional signal known as a “CAT box.” All of the signals necessary for expression of the nos gene are found within 261 bp of the nos transcriptional start site (Koncz, C. et al., supra). A gene for agrocinopine synthase and genes equivalent to tms and tmr have been identified on a nopaline-type plasmid (Joos, H. et al. [1983] Cell 32:1057-1067), and a number of transcripts have been mapped (Willmitzer, L. et al. [1983] Cell 32:1045-1056). McPhersson, J. C. et al. (1980) Proc. Natl. Acad. Sci. USA 77:2666-2670, reported the in vitro translation of T-DNA-encoded mRNAs from crown gall tissues.
Transcription from hairy root T-DNA has also been detected (Willmitzer, L. et al. [1982] Mol. Gen. Genet. 186:16-22). Functionally, the hairy root syndrome appears to be the equivalent of a crown gall tumor incited by a Ti plasmid mutated in tmr (White, F. F., E. W. Nester [1980] J. Bacteriol. 144:710-720).
In eukaryotes, methylation (especially of cytosine residues) of DNA is correlated with transcriptional inactivation; genes that are relatively under-methylated are transcribed into mRNA. Gelvin, S. B. et al. (1983) Nucl. Acids Res. 11:159-174, has found that the T-DNA in crown gall tumors is always present in at least one unmethylated copy. That the same genome may contain numerous other copies of T-DNA which are methylated suggests that the copies of T-DNA in excess of one may be biologically inert (see also Ooms, G. et al. [1982] Cell 30:589-597).
The Ti plasmid encodes other genes which are outside of the T-DNA region and are necessary for the infection process (see Holsters, M. et al. [1980] Plasmid 3:212-230, for nopaline plasmids; and De Greve, H. et al. [1981] Plasmid 6:235-248; Garfinkel, D. J., E. W. Nester [1980] J. Bacteriol. 144:732-743; and Ooms, G. [1980] J. Bacteriol. 144:82-91, for octopine plasmids). Most important are the one genes, which, when mutated, result in Ti plasmids incapable of oncogenicity. (These loci are also known as vir, for virulence.) Several one genes have been accurately mapped and have been found to be located in regions conserved among various Ti plasmids (Klee, H. J. et al. [1983] J. Bacteriol. 153:878-883; Iyer, V. N. et al. [1982] Mol. Gen. Genet. 188:418-424). The one genes function in trans, being capable of causing the transformation of plant cells with T-DNA of a different plasmid type and physically located on another plasmid (Hille, J. et al. [1982] Plasmid 7:107-118); Klee, H. J. et al. [1982] J. Bacteriol. 150:327-331; de Framond, A-J. et al. [1983] Biotechnol. 1:262-269). Nopaline Ti DNA has direct repeats of about 25 base pairs immediately adjacent to the left and right borders of the T-DNA which might be involved in either excision from the Ti plasmid or integration into the host genome (Yadav, N. S. et al. [1982] Proc. Natl. Acad. Sci. USA 79:6322-6326), and a homologous sequence has been observed adjacent to an octopine T-DNA border (Simpson, R. B. et al. [1982] Cell 29:1005-1014). Opine catabolism is specified by the occ and noc genes, respectively, of octopine- and nopaline-type plasmids. The Ti plasmid also encodes functions necessary for its own reproduction including an origin of replication. Ti plasmid transcripts have been detected in A. tumefaciens cells by Gelvin, S. B. et al. (1981) Plasmid 6:17-29, who found that T-DNA regions were weakly transcribed along with non-T-DNA sequences. Ti plasmid-determined characteristics have been reviewed by Merlo, supra (see especially Table II), and Ream and Gordon, supra.
Agrobacterium —TIP Plasmid DNA. Different octopine-type Ti plasmids are nearly 100% homologous to each other when examined by DNA hybridization (Currier, T. C., E. W. Nester [1976] J. Bacteriol. 126:157-165) or restriction enzyme analysis (Sciaky, D. et al. [1978] Plasmid 1:238-253). Nopaline-type Ti plasmids have as little as 67% homology to each other (Currier and Nester, supra). A survey revealed that different Ri plasmids are very homologous to each other (Costantino, P. et al. [1981] Plasmid 5:170-182). Drummond, N. H., M.-D. Chilton (1978) J. Bacteriol. 136:1178-1183, showed that proportionally small sections of octopine- and nopaline-type Ti plasmids were homologous to each other. These homologies were mapped in detail by Engler, G. et al. (1981) J. Mol. Biol. 152:183-208. They found that three of the four homologous regions were subdivided into three (overlapping the T-DNA), four (containing some one genes), and nine (having one genes) homologous sequences. The uninterrupted homology contains at least one tra gene (for conjugal transfer of the Ti plasmid to other bacterial cells), and genes involved in replication and incompatibility. This uninterrupted region has homology with a Sym plasmid (involved in symbiotic nitrogen fixation) from a species of Rhizobium, a different genus in the family Rhizobiaceae (Prakash, R. K. et al. [1982] Plasmid 7:271-280). The order of the four regions is not conserved, though they are all oriented in the same direction. Part of the T-DNA sequences is very highly conserved between nopaline and octopine plasmids (Chilton, M.-D. et al. [1978] Nature 275:147-149; Depicker, A et al. [1978] Nature 275:150-153). Ri plasmids have been shown to have extensive homology among themselves, and to both octopine (White, F. F., E. W. Nester [1980] J. Bacteriol. 144:710-720) and nopaline (Risuleo, G. et al. [1982] Plasmid 7:45-51) Ti plasmids, primarily in regions encoding one genes. Ri T-DNA contains extensive though weak homologies to T-DNA from both types of Ti plasmid (Willmitzer, L. et al. [1982] Mol. Gen. Genet. 186:16-22). Plant DNA from uninfected Nicotiana glauca contains sequences, referred to as cT-DNA (cellular T-DNA), that show homology to a portion of the Ri T-DNA (White, F. F. et al. [1983] Nature 301:348-350; Spano, L. et al. [1982] Plant Mol. Biol. 1:291-300). Huffman, G. A. et al. (1983) J. Bacteriol., have mapped the region of cross-hybridization and have shown that Ri plasmid, pRiA4b, is more closely related to a pTiA6 (octopine-type) than pTiT37 (nopaline-type) and that this Ri plasmid appears to carry sequence homologous to tms but not tmr. Their results also suggested that Ri T-DNA may be discontinuous, analogous to the case with octopine T-DNA.
It has been shown that a portion of the Ti (Chilton, M.-D. et al. [1977] Cell 11:263-271) or Ri (Chilton, M.-D. [1982] Nature 295:432-434; White, F. F. et al. [1982] Proc. Natl. Acad. Sci. USA 79:3193-3197; Willmitzer, L. [1982] Mol. Gen. Genet. 186:16-22) plasmid is found in the DNA of tumorous plant cells. The transferred DNA is known as T-DNA. T-DNA is integrated into the host DNA (Thomashow, M. F. et al. [1980] Proc. Natl. Acad. Sci. USA 77:6448-6452; Yadav, N. S. et al. [1980] Nature 287:458-461) in the nucleus (Nuti, M. P. et al. [1980] Plant Sci. Lett. 18:1-6; Willmitzer, L. et al. [1980] Nature 287:359-361; Chilton M.-D. et al. [1980] Proc. Natl. Acad. Sci. USA 77:4060-4064).
Thomashow, M. F. (1980) supra, and Thomashow, M. F. et al. (1980) Cell 19:729-739, found the T-DNA from octopine-type Ti plasmids to have been integrated in two separate sections, TL-DNA and TR-DNA, left and right T-DNAs, respectively. The copy numbers of TR and TL can vary (Merlo, D. J. et al. [1980] Mol. Gen. Genet. 177:637-643). A core of T-DNA is highly homologous to nopaline T-DNA (Chilton et al. [1978] supra; and Depicker et al. [1978] supra), is required for tumor maintenance, is found in TL, is generally present in one copy per cell, and codes for the genes tms, tmr, and tml. On the other hand, TR can be totally dispensed with (De Beuckeleer, M. et al. [1981] Mol. Gen. Genet. 183:283-288; Ooms, G. et al. [1982] Cell 30:589-597), though found in a high copy number (Merlo et al. [1980] supra). Ooms, G. et al. (1982) Plasmid 7:15-29, hypothesized that TR is involved in T-DNA integration, though they find that when TR is deleted from the Ti plasmid, A. tumefaciens does retain some virulence. Ooms, G. et al. (1982) Cell, supra, showed that though T-DNA is occasionally deleted after integration in the plant genome, it is generally stable and that tumors containing a mixture of cells that differ in T-DNA organization are the result of multiple transformation events. The ocs is found in TL but can be deleted from the plant genome without loss of phenotypes related to tumorous growth. The left border of integrated TL has been observed to be composed of repeats of T-DNA sequences which are in either direct or inverted orientations (Simpson, R. B. et al. [1982] Cell 29:1005-1014).
In contrast to the situation in octopine-type tumors, nopaline T-DNA is integrated into the host genome in one continuous fragment (Lemmers, M. et al. [1980] J. Mol. Biol. 144:353-376; Zambryski, P. et al. [1980] Science 209:1385-1391). Direct tandem repeats were observed. T-DNA of plants regenerated from teratomas had minor modifications in the border fragments of the inserted DNA (Lemmers et al., supra). Sequence analysis of the junction between the right and left borders revealed a number of direct repeats and one inverted repeat. The latter spanned the junction (Zambryski et al. [1980] supra). The left junction has been shown to vary by at least 70 base pairs while the right junction varies no more than a single nucleotide (Zambryski, P. et al. [1982] J. Mol. Appl. Genet. 1:361-370). Left and right borders in junctions of tandem arrays were separated by spacers which could be over 130 bp. The spacers were of unknown origin and contained some T-DNA sequences. T-DNA was found to be integrated into both repeated and low copy number host sequences. H. Joos et al. (1983) Cell 32:1057-1067, have shown that virulence is not eliminated after deletion of either of the usual nopaline T-DNA borders.
Simpson et al. (1982) supra, and Zambryski et al. (1980) supra, have suggested that direct repeats in the border regions are involved in integration of T-DNA into plant DNA. That T-DNA having borders from two different Ti plasmids are less specifically integrated than are homologous borders supports this suggestion (Ooms, G. et al. [1982] Plant Mol. Biol. 1:265-276).
Yadav, N. S. et al. (1982) Proc. Natl. Acad. Sci. USA 79:6322-6326, have found a chi site, which in the bacteriophage (Greek symbol) augments general recombination in the surrounding DNA as far as 10 kilobases away, in a nopaline Ti plasmid just outside the left end of the T-DNA. Simpson, R. B. et al. (1982) Cell 29:1005-1014, have not observed a chi sequence in an octopine Ti plasmid, though the possible range of action does not eliminate the possibility of one being necessary and present but outside of the region sequenced. The significance of the chi in the Ti plasmid is not known. If the chi has a function, it is probably used in Agrobacterium cells and not in the plants, as chi is not found within the T-DNA.
Agrobacterium—Manipulations of the TIP Plasmids. As detailed in the section on Shuttle Vectors, technology has been developed for the introduction of altered DNA sequences into desired locations on a TIP plasmid. Transposons can be easily inserted using this technology (Garfinkel, D. J. et al. [1981] Cell 27:143-153). Hernalsteen, J.-P. et al. (1980) Nature 287:645-656, have shown that a DNA sequence (here a bacterial transposon) inserted into T-DNA in the Ti plasmid is transferred and integrated into the recipient plant's genome. Though insertion of foreign DNA has been done with a number of genes from different sources, to date foreign genes have not usually been expressed under control of their own promoters. Sources of these genes include alcohol dehydrogenase (Adh) from yeast (Barton, K. A. et al. [1983] Cell 32:1033-1043), AdhI (Bennetzen, J., unpublished) and zein from corn, interferon and globin from mammals, and the mammalian virus SV40 (Schell, J., unpublished). However, when the nopaline synthase gene was inserted into octopine T-DNA and transformed into plant tissue, it was found to be fully functional (Fink, C. L. [1982] M. S. thesis, University of Wisconsin-Madison). The gene encoding phaseolin, the storage protein found in seeds of the bean Phaseolus vulgaris L., has been transferred into and expressed in sunflower tumors. This latter work constitutes the first example of a transferred plant gene being expressed under control of its own promoter in foreign plant tissue. Transcription started and stopped at the correct positions, and introns were post-transcriptionally processed properly (Hall, T. C. et al., U.S. application Ser. No. 06/485,613, which is hereby incorporated by reference). Hosters, M. et al. (1982) Mol. Gen. Genet. 185:283-289, have shown that a bacterial transposon (Tn7) inserted into T-DNA could be recovered in a fully functional and seemingly unchanged form after integration into a plant genome.
Deletions can be generated in a TIP plasmid by several methods. Shuttle vectors can be used to introduce deletions constructed by standard recombinant DNA techniques (Cohen and Boyer, U.S. Pat. No. 4,237,224). Deletions with one predetermined end can be created by the improper excision of transposons (Koekman, B. P. et al. [1979] Plasmid 2:343-357, and Ooms, G. et al. [1982] Plasmid 7:15-29). Hille, J., R. Schilperoot (1981) Plasmid 6:151-154, have demonstrated that deletions having both ends at predetermined positions can be generated by the use of two transposons. The technique can also be used to construct “recombinant DNA” molecules in vivo.
The nopaline synthase gene has been used for insertion of DNA segments coding for drug resistance that can be used to select for transformed plant cells. In plant cells, the kanamycin resistance gene from Tn5 is not transcribed under control of its own promoter (Kemp, J. D. et al. [18 May 1982] Beltsville Symp. VII, Beltsville, Md. [1983] in Genetic Engineering: Applications to Agriculture, Owens, L. D. (ed.); and Fink, C. L. [1982] supra). Bevan, M. W. et al. (1983) Nature 340:184-187 and Fraley, R. T. et al. (1983) Proc. Natl. Acad. Sci. USA 80:4803-4807, have inserted the kanamycin resistance gene (neomycin phosphotransferase II) from Tn5 behind (i.e., under control of) the nopaline promoter. The construction was used to transform plant cells which in culture displayed resistance to kanamycin and its analogs such as G418. Schell, J. et al. (18 Jan. 1983) 15th Miami Winter Symp. (see also Marx, J. L. [1983] Science 219:830), reported a similar construction, in which the methotrexate resistance gene (dihydrofolate reductase) from Tn7 was placed behind the nopaline synthase promoter. Transformed cells were resistant to methotrexate. Similarly, Hererra-Estrella, L. et al. (1983) Nature 303:209-213, have obtained expression in plant cells of enzymatic activity for octopine synthase and chloramphenicol acetyltransferase, an enzyme which in bacteria confers resistance to chloramphenicol, by placing the structural genes for these two enzymes under control of nos promoters.
Hall, T. C. et al., U.S. application Ser. No. 06/485,614, which is hereby incorporated by reference, have fused the ocs promoter and the 5′ end of the octopine synthase structural gene to the structural gene for the bean seed protein phaseolin. A fusion protein having the amino terminus of octopine synthase and lacking the amino terminus of phaseolin was produced under control of the T-DNA promoter. The introns, which were contributed by the phaseolin sequences, were post-transcriptionally processed properly.
de Framond, A. J. et al. (1983) Biotechnol. 1:262-269, have reported on the construction of a “mini-Ti plasmid.” In the nopaline T-DNA there is normally only one site cut by the restriction enzyme KpnI. A mutant lacking the site was constructed and a KpnI fragment, containing the entire nopaline T-DNA, was isolated. This fragment together with a kanamycin resistance gene was inserted into pRK290, thereby resulting in a plasmid which could be maintained in A. tumefaciens and lacked almost all non-T-DNA Ti sequences. By itself, this plasmid was not able to transform plant cells. However, when placed in an A. tumefaciens strain containing an octopine Ti plasmid, tumors were induced which synthesized both octopine and nopaline. The mini-Ti plasmid has also been transferred into plant cells when complemented with a Ti plasmid deleted for its own T-DNA. These results indicated that the non-T-DNA functions acted in trans with T-DNA, that the missing nopaline Ti plasmid functions were complemented by the octopine Ti plasmid, and that the nopaline “mini-Ti” was functional in the transformation of plant cells. A similar pair of complementing plasmids, each containing either octopine T-DNA or one genes, has been constructed by Hoekema, A. et al. (1983) Nature 303:179-180.
Chilton et al. (18 Jan. 1983) 15th Miami Winter Symp., also reported on the construction of a “micro-Ti” plasmid made by resectioning the mini-Ti with SmaI to delete essentially all of T-DNA but the nopaline synthase gene and the left and right borders. The micro-Ti was inserted into a modified pRK290 plasmid that was missing its SmaI site, and was employed in a manner similar to mini-Ti, with comparable results.